#64 Matt McIlwain – MD Madrona Venture Group on all Things Venture Capital

Hello everyone, welcome to another episode of Startup Project where we talk about the business of technology, startups, and venture capital. In today's episode, we have Matt from Madrona Ventures. He's a managing director at Madrona Venture Group, which is an iconic venture firm here in Seattle. Um, they funded companies like Amazon, AppTio, Redfin, Snowflake, and UiPath. Um, Before joining Medrona, Matt worked at McKinsey and was a VP of Business Process at Genuine Parts Company. He's a graduate from Dartmouth and has an MBA from Harvard. In this episode, we discuss, you know, all about venture capital, investing in great companies, and what Matt is seeing in what is happening in venture capitals broadly and specifically in Seattle area. I hope you enjoy this conversation. Hey, Matt, it's great to have you with us. Welcome to the show.

Thank you very much.

So to just set context for our audience, obviously you work now at Medtrona Venture Group, but can you talk a little bit about your early career before Medtrona and what you've done before?

Well, sure. I did a mix of things prior to Medrona. I was in investment banking right out of college and then went back to graduate school. And in graduate school, I had the opportunity to go to work for McKinsey and Company. And so spent 4.5 years at McKinsey doing strategy and management consulting, particularly in technology-driven industries. That led me to an opportunity in a holding company which had a bunch of different businesses. And those businesses in the mid-'90s were getting disrupted by this thing called the internet. And so I had this opportunity to work both inside of our businesses, and this was things like Napa Auto Parts and industrial products, like really traditional businesses, and figure out how to embrace all the technology trends and capabilities as well as some of the business model implications. Including working with a whole bunch of companies that were venture-backed startups. And that's how I got to know even more deeply the venture-backed startup world, which eventually took me to joining Madrona back in 2000. And as you know, you know, Madrona had been the first investors in Amazon. And so one of the ways I got to know the Madrona team was exploring a joint venture that we ended up not doing between Madrona and Napa Auto Parts, and Amazon.

So you've been at Madrona, if I'm not wrong, 20-plus years.

Yeah, it's approaching 24.

So obviously you've seen multiple venture cycles, right? I mean, we've seen that 2008 crash and you could probably say the 2015 correction, right? You've been around to see the ups and downs of this, you know, venture and in general, the markets in general. Where do you see the venture landscape today?

Yeah, well, I do think it's helpful to have had that perspective of these different market cycles. You know, I joined right at the peak at the time of the Nasdaq index in, you know, kind of early 2000, you know, kind of April 2000 and actually joined full-time in May. And then right away, the market started to have a big downturn. And so you see then in those kinds of environments how important it is to react smartly to the market conditions around you. You know, we always think at Medtrona about the things you can control and the things you can't control. And the things you can't control, you need to deeply understand because they should influence at some level, you know, your operating plan and your focus on particular use cases and how many different experiments you can be running at once. And so to get to your specific question, I think we're in an environment right now, which is a bit of a, you know, kind of a tale of two cities, as it were, or two barbells, depending on how you want to think about it. You know, on the one hand, there is a lot of businesses that are challenged. They're later stage businesses in many cases, and they have not fully found a scalable product market fit. They may have found product market fit, but they haven't found scalable capital efficient product market fit. And it's a particularly tough time in the later stage world of both investing in companies as well as, you know, M&A transactions. But on the other hand, there are a number of companies that had been either more disciplined on the cash side or got started a little bit later, a little bit outside of the sort of the euphoria of, you know, maybe that, you know, 2019 to 2021 time period. And those companies are doing quite well, particularly those that have embraced and are really kind of GenAI native. Style of companies. So they're having a little trouble raising capital. They're having good success at working with different kinds of partners and building products and, you know, good at growing their businesses. So it's a little bit of a bifurcated time.

So there's this sort of like disconnect in terms of like how public market is viewing private companies versus how the private market is viewing And how do you see sort of this difference in between like how the public is judging startups when they go public versus we're still seeing some high valuations, right? Even at late stage, like how do you see that disconnect? Or if there is one.

I think that one of the big differences between the public and private markets is the public markets The value of companies is changed and assessed every day, basically, especially with the activeness of the aftermarkets versus the kind of the common public market trading windows that take place day in and day out. And so, you know, those markets went way up. And, you know, one of the best metrics I like to use is the average revenue multiple for public software as a service companies. And so if you look at like the last 10 years, the average revenue multiple, you know, forward 12-month revenue multiple is about 7 times. In other words, if I had a company and they did a bill, they're expecting to do $1 billion in revenue over the next year, then they would be valued around $7 billion on average today. For kind of a reasonable growth rate. And then there was an era where that average, you know, for a period of time, a short period of time in 2021, went as high as 15 times. So that same company with that same billion of revenue and same growth rate would be worth $15 billion rather than $7 billion. And now that's come back down in the public markets to $5 billion or 5x. So we've gone from 7x to 15x back down to 5x. All in the public markets, all within 2 years. The private markets are much slower to adjust to those types of things because you only really adjust the value of a company's shares when, you know, formally when it raises a new financing round. And it often raises those at higher valuations and sometimes at lower valuations. And so there's many, many companies that raised money in 2021 as private companies and are still private, and they haven't yet actually raised a new round since then. So while some people either in the secondary markets or in their own private books of how they keep track of these companies may have lowered their valuations, their official valuations and their certainly their mindset around what they're valued at hasn't changed all that much in many cases. And that kind of highlights the difficulty in the disconnect between the public markets and the private markets. And so I guess a related question or two related questions to that then would be, Oh, if I'm thinking about going public and I raised money in 2021, and at that time I was able to command a, you know, 15 times revenue multiple, and maybe I was doing, you know, forward, you know, revenue of $500 million. Well, you know, then that would be a company that's worth $7.5 billion back in 2021. But now at a 5x forward revenue multiple, I might only be worth $2.5 billion. Do I want to go public if that's what the market, the public market's going to value me at? Or am I willing to stay private? Can I stay private longer because I have enough cash and enough of a cash-generating business to remain private and independent until I can, quote, grow into some of those, you know, 2021 levels of valuations by growing my revenue?

What do you think about this idea of, you know, some of the funds have seen raising just purely to invest in late-stage companies, sort of like what Tiger Global and SoftBank did, but in next, I don't know, 3, 4 years? Like, do you think that might be a better strategy? Or like, what do you think about that strategy in general?

Well, I think if the question is, you know, is there going to be a re-emerging market for later-stage growth rounds? In the private markets, I think the answer is that time will come. And I think there's a few pieces at play here. So one question you'd have to be kind of thinking through is, you know, will some of these public investors that came into the private market come back into the private markets? A lot of them have gone away from the private markets. That's probably positive from the perspective of another investor, like in your example, somebody new coming in with a new fund because it's not as crowded a market. Crowded markets with lots of demand for something tend to drive up prices. So that's one consideration. The second consideration is, you know, what happens in the public market valuations and how does that sort of cascade down into private market valuations? So could I get into later stage companies at comparatively better or lower private valuations because they've eventually caught up with the public valuations and there's enough new companies that are starting to move their way into the growth stage that I might get some quote better, you know, kind of value for my dollar. The third thing is, and we've certainly seen some of this in the last, you know, 12 to 18 months is sometimes you can go invest in a company not by buying new shares they issue, which is a primary round, but you buy existing shares from some other owner of those shares in a secondary purchase. And oftentimes you can get those secondary purchases at a lower price, at a significant discount to the last primary round. Part of the reason is, is because you're stretching out these periods where there's not really great liquidity in that company's stock. And so somebody's willing to sell at a discount because they want to be liquid. They want to sell now. And so I think that's the place that's probably the hottest at the moment in the later stage companies is secondary purchases in their shares, whether that's from an early investor who for them, maybe they're still making good money on that and on that sale or a founder or longtime employees. And there's a whole bunch of details we could dig into on that front. But I think those are some of the ways that it might be interesting to explore and potentially invest in later stage companies.

I think the difference is in the last 2-3 years, even getting secondary transactions done was difficult because all the companies were buying back their own secondary shares when they had right of first refusal. So we would have seen most of those transactions at higher price. And in some cases they've been blocked by the company because the company was purchasing their own shares, which is slightly different from now.

Well, the right of first refusal is, is a mechanism, and some companies have chosen to do a bit of that. I don't think that's been all that pervasive, partly because I think fairly quickly after things turned around in late 2021, a lot of these companies were focused on kind of careful cash management. And so if they were going to buy those shares back, they were going to be forced to buy them back at the price that somebody was able to get in the other secondary markets. And that may not be the best use of their cash. But the other mechanism that companies could do is in some different ways they could either have explicit restrictions or create kind of ground rules around we're only going to allow you to sell so many of your shares. And so that creates some complexity for employees. There are some workarounds to that. Again, there's lots of second and third order details here you could get into. But yes, there's been some roadblocks to doing some of those secondary transactions, but there have been people that have found ways and willing sellers to actually get positions in some of those companies.

I feel like sort of venture came into the retail landscape a little bit in terms of like it became more popular in the last 3, 4 years. And, you know, with IPOs and SPACs, because startups were so inaccessible and stayed so private and overall felt like there was a lot more attention to like VCs and, you know, what investments are being made into startups, etc., right? Did that attention change in terms of like LP perspective? Like, are LPs allocating same amounts to venture capital? Like, are new funds able to raise in a similar fashion? Like, did that change?

Well, that's a very interesting question on the limited partners. The folks that invest into venture capital firms and how their mindset has shifted. You know, one of the positive developments that, you know, kind of came about starting about 5 years ago, kind of 2018 timeframe through 2021, early 2022, is a lot of companies were going public, a lot of companies were being acquired. And so the funds were returning significant amounts of capital to their LPs. And so the LPs were, getting liquidity, they were getting returns. And so I think that actually led to a season of those LPs being more open and willing to invest in the next fund. You know, I've been successful with this venture firm, I'm going to invest in their next fund. And I think the pace of that got too quick and the size of many of those funds got too big, candidly. And so then, you know, as a result of that, you also saw the venture funds investing much more aggressively. And just to quantify that, and I'm going to use round numbers here, you know, up until 2021, the biggest year ever in history of venture capital dollars in the US going in was about $150 billion. 2021 was this huge jump up to $350 billion. And then last year in 2022, it dropped to $250. So now let's just pause on that a second. So the $250 billion in 2022 is actually the second largest by far year in the history of investing in companies. You know, this is again the venture firms investing in companies. And even though this year it's going to drop again by another $100-ish billion back down to about $150 billion, it's still going to be roughly tied for third in terms of the all-time biggest year in venture dollars going So while we've had this massive correction relative to 2021 in venture money being put to work, we're still at, you know, kind of at and approaching historic all-time highs on an annual basis. Now, when you step back and say, well, what's happening with the limited partners, the university endowments and foundations and pension funds and what are called sovereign wealth funds or the funds, the endowments of big country nation states and small country nation states for that matter, they're sovereign funds. I think there's been some introspection there and stepping back and analyzing, well, what did we not like? We didn't like the pace of new fundraisers. We now don't like all these write-downs that are happening in companies. We also don't like the fact that there hasn't been material returns back to us. There hasn't been good liquidity. And I think they are slowing their pace and doing really two things just to summarize it. One is to make less new fund commitments. So it's harder to get some new endowment to invest in me if they hadn't previously been an investment. And then secondly, they're even paring back, either actually not backing some of the firms they backed in the past or putting less money into them, or maybe the same money as they did in the last, in the last fund versus expanding it. So I think that's some of the dynamic. You're really going to see that into 2024. 2023, I think, has been just a tough year in general for venture fundraising. You know, speaking of Madrona, we raised our new fund in late 2022, and that's about a $700 million fund. It's actually two different funds of seed and Series A fund, and then a kind of a Series B, Series C fund, which we call our acceleration fund. And so we're fortunate to be in a very strong position in terms of the capital, you know, that we can invest in, you know, in making a number of investments in new companies.

I'm glad you brought that up because the size of the fund sort of drastically changed in the last 5 years, right? We've seen funds with billion dollars, you know, multi-billion dollar size funds in crypto. And you've consistently kept that fund size, I would say, you know, in looking at the industry, sort of conservative. Like I think Benchmark also in its recent fund was around $600 million, if I'm not wrong. How do you see the fund size approach? Like, is there a point where if you raise beyond a certain size, does it become much harder to return that fund?

Well, I think that, you know, it is harder to return a fund if the fund size gets bigger. And we can talk a little bit about the math around that. And, you know, we've always had a practice of aspiring to have every dollar that our investors entrust us with that we can somehow over time turn that into $5. So 5x return. And we also make a big point of being aligned with them by putting a substantial portion of our own capital into every one of our funds. And we've always done that and we always will. And so those two things are a bit of sort of governors on as you really think through, Okay, you know, what fund sizes do I want to have? What fund sizes are aligned with my longstanding investor partners in terms of our collective desire to kind of generate those, you know, industry-leading, you know, styles of return? You know, in reality, if you can generate a 3x cash-on-cash to turn, you know, $700 million, to use round numbers in our case, into $2.1 billion over the course of time, you're going to be a top quartile, if not top decile performer in every fund cycle, every vintage year as they're called. But then you say, well, let's say if you had, you know, 25 companies that you invested in out of a $700 million fund, you know, and I'm using just round numbers here, but you know, you would need to have, you know, a few of those companies on their own return the entire fund in order to have really any shot at getting to the $2 billion plus in returns. And that's hard to do because especially if I'm now at a $700 million fund in aggregate, we technically have these two funds, but I'm using these simple numbers. That means that some company, let's say you own 20% of it, has to be worth $3.5 billion when it gets sold or it goes public in order for your 20% to be worth $700 million. And there's not too many companies that end up worth $3.5 billion and not just worth it on paper, but worth it in an actual monetization event, whether that's an M&A transaction or sustainably post-IPO. And so it is really hard. And again, you probably need 2 of those in a fund and then you need some other pretty good results on top of that to get to 3x or better. So a lot of work to be put into and it only gets harder when you raise bigger funds.

So talking about, you know, you mentioned investing in 25 companies for the fund, but you also mentioned like you have the seed and Series A and your growth stage fund, right? Like how are you thinking about— I think we've seen this phenomenon as well where venture funds going early in general, like there used to be a much more clear distinction of a venture fund or firm being, you know, we are a Series A firm, we're a Series B firm. Well, we've seen this new approach where firms are also doing seed plus Series A or, you know, multiple stages. Like how What do you at Madrona think about that?

Well, our core strategy for a long time, really since the beginning of the firm, has been being there at an early stage with companies. And so we love investing in pre-seed rounds, seed rounds, Series A rounds. And then we have this expression of being there at day one and being there for the long run. So this idea that not just with our capital, because even if we invest $5 million or whatever millions of small millions of dollars in the first investment, We could end up putting $30-40 million into an individual company over the life of the company. And so there's this capital element of being there and reserving capital in a fund to support a company all the way through the journey. But more importantly is rolling up our sleeves as the partner who's on the board, the team at Madrona, that's the main team, as well as the broader Madrona team and family that are doing everything they can to help you as a founder and your team win over the long term. So that's our bread and butter. That's what we love. And I think back to companies like Isilon Systems that I helped, you know, from day one in the Series A round back in 2001. That was back in the day when a Series A round was the first round in many cases. And then you had Smartsheet and Appio in 2007, or I could think about Placed or Turi. I mean, these are all different examples of companies that we were there you know, at literally day one and help them build those companies over the long term. In fact, you know, 15+ years later, I'm still on the board of Smartsheet, which is a public company now. So then you've also got to say, well, about 10 years ago, we realized, especially since a lot of our investments, not all of them, but many of them are based in the Pacific Northwest, we said, gosh, once a company's found product market fit and they're in some of these sectors that Because of our focus in Seattle, we deeply understand sectors like software as a service, sectors like cloud, sectors like applied AI, that in those areas there might be some great companies that are outside of the region that we could add a lot of value to. And we also have some uniquely deep and special relationships with Microsoft and Amazon, and they're based in Seattle too. So why don't we go try to help a handful of those companies? And that's what led to this other vehicle, which we call the Acceleration Fund, where it's a new investment for us, usually Series B, maybe Series C stage, broader geographic look. And we, we're gonna roll up our sleeves and help those companies too, just getting in a little bit later. And happily, we made investments in companies like Snowflake and Accolade and UiPath and Go One in that strategy. And that's worked very, very well for us and for those companies where we've added complementary value. So that's kind of how we think about it. There's this seed and Series A focused fund, and then there's this kind of acceleration Series B, C fund with a little broader geographic footprint. And both of those have been quite successful for us. And hopefully as we work hard, we will keep being successful.

So you mentioned Pacific Northwest, and obviously Medrona is located in Seattle. What do you see is happening in Seattle ecosystem? We talked about Broadly Venture, but do you see some parts missing? You know, what's working, what's not working in Seattle?

Well, I think some of the great things about Seattle are that there is just an incredible amount of talent and talent that's in the flow of these big sector changes well before most of the world, you know, is aware of it, you know, you know, the Amazon Cloud back in 2006, 2007, you know, relatively quickly followed by Azure over at Microsoft and then applied AI and ML. I mean, we made our first AI investment in 2013 and it was the Amazon Professor of Machine Learning at the University of Washington and his grad students, Carlos Guestrin, that company Turi ended up getting bought by Apple. So I think we've got the talented people, they're in the flow of where innovation is coming. And they've got great experiences. I think that we are a bit underfunded, honestly, from a venture capital perspective. You know, we're the largest firm in the region and we're super hungry and know we got to earn that every day. And again, we're really delighted to get involved at a pre-seed stage, but we love having great collaborators. I think far too often a lot of our collaborators have come from outside of the region, and we welcome and love working with the different firms, you know, the, you know, the kind of the big names, you know, the Sequoias and Kleiners and Greylocks and Andreessen and others. We've worked with all those folks. They're terrific. But having more of the kind of local capital, especially early on, I think is really, really important to collaborate and help build, you know, with founders together from an early stage. That's probably the thing that's most missing. I think the other thing is that, you know, there's been different sort of waves and eras of people coming out of the Microsofts and Amazons and Googles skills and being able to kind of adjust well to working in a much smaller environment. I think there's a number of folks that have been able to do that. You know, you think of, you know, some of our recent investments like, you know, Abe, who was at Microsoft for many years, is the CEO at Typeface, or Jordan Tagani over at Motherduck, which is a terrific company. You know, there's people that have come out of Apple now, you know, and certainly Amazon as well. So, I think it's gotten better, but I think there's still some more opportunity there with those broader ecosystems around Microsoft, Amazon, Google, and others.

So, Madrona, you know, I think this is the second time you've started doing the Intelligent Application Summit. And can you talk about— and I was going through, you know, the guest list and you have really amazing guests, you know, founder of LangChain, Brad Gersner from Ultimator. And some amazing folks from Microsoft, Amazon. I think AWS CEO was there. So, and I was going to some of the conversations and they're really fascinating. Can you talk a little bit about, you know, what was the summit about and what are some of the key takeaways?

Well, thanks. And that, yeah, that was, it was a great summit and we will be putting out most of the different, you know, panels and keynotes and videos. And maybe you wanna link to in the show notes to, Madrona, where we, we'll have a bunch of those. And on the I-40 site. And it goes back a bit to the fact that we had, you know, we had started investing in this area over a decade ago and have built some great relationships over time. In fact, if you actually go way, way back, Brad Gerstner and I were on the board almost 20 years ago of a company called FairCast that was founded by Oren Etzioni, who was then a professor and was doing some early things and kind of scalable statistical, you know, analysis and machine learning that predicted whether or not airfares were going to go up or down in the future. That company ended up getting bought by Microsoft and becoming Bing Travel. So there's a long history of using data, using statistical methods, machine learning, now, of course, foundation model, you know, based techniques to innovate and build, you know, build the building blocks and then build the actual intelligent apps themselves. So about 3 years ago, we invited the venture community to help us identify what are the most promising companies, this Intelligent Application 40. And there were nominations. We worked with our friends at PitchBook, and we ultimately had a voting mechanism that led to this selection of 40 companies. There's some early, some mid, some late-stage intelligent app companies and some enablers. What's been interesting now in the 3 years that we've done that, you know, process in that sort of selection of the Intelligent Application 40, and these are all private companies. There's only 7 companies that have won all 3 years. And I think that speaks to one of the insights, which is this is a very dynamic world right now, this world of applied AI, the pace at which innovation is happening, the pace at which some kinds, some kind of companies are starting to break out, but then hit some roadblocks and need to navigate through that. Now, specific, in some specific areas that we think that are quite interesting. One is that a point of view that we had a while back that I think has now become more fully adopted is it's going to be a mini-model world. There's going to be many models, not few models or one model. And those models are often going to be used and combined. We had Ali Farhadi, who actually just succeeded Oren Etzioni as the head of the Artificial Intelligence Institute, come and give a talk. He talked about this, you know, technical concept of ensemble models. In addition to model ensembling, there's just going to be model combining, or what I like to call model cocktails. And those models will come together in ways that help you make and build better intelligent applications. So that was one of the big discussion topics, one of the big takeaways. Another is that a lot of people are in fact experimenting with generative AI and foundation models, but they're more at the prototype and experimentation stage. Than they really are in production. So we're still early in the journey from prototype to production. And then the third area that I would, I would highlight is that there really is going to be this interesting sort of evolution. And you could think about it as a bit of a battle, as it were, from a competitive perspective of the incumbent software companies that are gen, you know, enhancing their applications with AI capabilities and the true gen-native companies that are from first principles building a gen-native foundation model derived application such as a Typeface, such as a RunwayML, both of those being two of the 40 winners. And so in contrast to the Adobes and even, you know, Microsofts and Salesforces, and it looks like the incumbents are actually going to succeed quite well in the short term because they have the customer relationships, they have the data, they've been quick to embrace this technological disruption. We believe that there's going to be some big winners that are Gen Native, but it's probably going to take more of a medium-term time horizon, a 3 to 5 year time horizon for it to be fully evident which ones of those are going to be the bigger winners. So those are just some of the topics from the summit, but happy to talk about more and any other specific questions.

You brought up the point of incumbents versus startups, and one of the takeaways for me or opinion of me is it used to be very difficult for big companies to sort of adapt to new things. And I think that sort of changed from what we're seeing from the AI cycle to sort of previous cycles, like even mobile or previous shifts, technology shifts that we've seen, is big companies are actually also agile. And one of the reasons we are seeing this AI adoption into big companies is because at the end of the day, you're providing AI as a sort of like an API as a service. And when you combine these two things, that the big tech companies are actually agile because the individual teams inside them are agile, and when the technology is available as an API, and that's sort of like made it much easier for even larger companies to adopt and create solutions as fast as we are seeing. So that's sort of like my takeaway in terms of like why we are seeing this incumbents being so fast to this. And I'm curious to get, you know, your thoughts on that.

Well, I think the API point is a good point. And I would add in a couple of other areas. You know, one is that, you know, they do have an immense— these incumbents have an immense amount of data around which they can, you know, train models or fine-tune models, or in some cases bring in augmented data in the form of, you know, some kind of, you know, a system using a vector database and something like retrieval augmented generation. And they have that data. They, and they also have context about what their customers are trying to do. And so I think you've got some incumbent advantages. Now you've also got, you know, an existing, you know, relationship where your customer is already using your software. And if I can enhance that software with some of these features and capabilities, then I get more engagement. That's what we could sometimes refer to as sort of the reinforcement learning from human feedback within my now enhanced application that I pull back into my models and continually make my models better. So I think there's a bull case for the incumbents. The bear case for the incumbents is, is that, you know, historically, software was run very deterministically. It was very, you know, predictable. It ran the same way basically every time. Sure, it was configured and customized a little bit, but it basically basically runs the same way. It's meant to be that way. And by definition, these, you know, types of generative models are running predictively. They're running non-deterministically. And so there's going to be some amount of uncertainty in what results they come back with. That might be unsettling for the end user customer of a big company. And they might have different expectations about that. It could create some friction, some tension for the incumbents. I think there's also some complexity around pricing and packaging here. There's a lot of early experimentation going on. And then finally, just, you know, well, I think you're right. It is somewhat different than the mobile era. You know, there were a lot of companies, for instance, in travel that built a mobile app eventually, but then they didn't invent Airbnb. They didn't invent Uber or Lyft. So who are going to be the companies that from a first principles perspective just reinvent something? you know, that solves a customer's problem better than it's been solved before in categories. And for that matter, in travel or consumer and consumer shopping or, or other aspects of how enterprises get work done. So I think there's, again, it's gonna take a while because they don't have some of those incumbent advantages, but I think you're gonna see some winners emerge over the, you know, kind of the medium term.

I mean, even the whole AI boom, if you, if you can call that, started with OpenAI, which is, not an incumbent, right? And another thing probably that is good for non-incumbents is like if you think of incumbents like Microsoft and if you look at a product like Office 365 or Word, right, you're sort of trying to add, okay, now we have AI and how do we make this better? But maybe AI will now allow you to create new form factors, right? You don't have to have a note-taking tool similar to Word, right? You might have to think it a little bit different. Same with like if you have a video editing tool, AI does most of the work, does it have to be how Adobe's product is going to be or something completely new, right? I think that's where like startups have the edge on redefining some of these products because AI enables you to do those new things.

And yeah, I think, I think that's right. And, and, and then, and I think that there is this sort of what does the technology enable? And then it also is like, what does it enable from a business model perspective? So one of my favorite examples to illustrate this is, is that the world today of search and shopping digitally online is largely a seller's agent world. There's some brand that's trying to sell something to me as the customer. And there's a toll taker, whether their name is Facebook or Google that take or Amazon that takes a large amount of the toll, you know, from that seller trying to reach me. And I think in this world of very personalized and customized AI, to the extent I trust a buyer's agent, the buyer's agent who represents me as the buyer could completely change not only the equation of how I go do my shopping, but also how the money flows, how the business model works for consumer commerce. Now, I haven't seen the example of that figured out. Directly yet. And honestly, I think that, you know, people have tried it before, but this is one of these areas where potentially generative AI capabilities and what's coming next beyond just purely generative that we think there are some interesting things happening back to my example of ensemble models and what's happening in, you know, can even in the model ecosystem. I think there's a lot, you know, to be said about different kinds of end-user empowered agents versus services that are more, you know, kind of bought, you know, kind of pushed through the system from the, from the supplier or the seller side.

Yeah, I mean, I really like that idea of like having your own agent go and shop for what you want and do it in an intelligent way, just like you're being recommended something on Instagram, right? Like, that's an interesting idea. I haven't seen yet someone do it, but I think it would be a great startup idea for sure. So we're almost at the end of our conversation, but I wanted to ask one more question around the AI. What do you think of this strategy that sort of like all the big tech companies seems to have adopted? Microsoft being partnered with OpenAI and Amazon seems to be partnering with Anthropic. And I think Oracle announced some partnership, but I forgot which company. But we are seeing that The incumbent companies are partnering with large language model companies or foundational model companies. Is it purely about because these companies have cloud as a product and OpenAI or Anthropic, they can use these resources, which is I think a short-term supply restraint? Is that the main reason or what do you think is really playing out with the strategy?

Well, I think there could be some different reasons here. You know, there's other companies like Cohere. There's yet others that are emerging, and there's different other kinds of models like Falcon. So there's a growing number again of these models. I think some of the most, you know, talented and early teams have partnered up with somebody, you know, and you mentioned the OpenAI Microsoft relationship. You know, interesting fact, when OpenAI was still a nonprofit in the early years, Amazon was actually a contributor and Microsoft was not in the very early days. So, you know, I think these things have evolved a bit. What I think is more notable, though, is that even the Microsofts and Amazons with their partnerships, and I would probably say in this case, especially Amazon has taken a more open approach and said, well, yes, you can use, you know, and we have now they've announced, you know, this sort of special relationship with Anthropic. But they've also got the Llama 2 models that are Meta's models. They've also got, you know, available on a service they call Bedrock. They've also got, you know, Stable Diffusion's models, you know, the Stability AI models that are available there too. So I think they're taking more of a marketplace approach, maybe not surprising for Amazon. And even Microsoft has also said, hey, you can get Llama 2 available as a service. And I think that those groups in different forms and fashion will have some of their own models available as a service too. Clearly Google does, you know, we're all waiting to see what Gemini is all about. And in the case of Google, they also have their own, you know, training, you know, apps or rather training, you know, chips. And so that's a whole other set of questions around the NVIDIA, you know, ecosystem and how NVIDIA competes and complements and works together with the cloud service providers where you've got, you know, tensor processing units at Google, you've got, You know, for Amazon, again, you've got both a training specialty chipset called Trainium and then something called Inferentia for inference. And Microsoft broadly rumored, and I think we'll hear something more from them within the next month about their chips that they're building and who they're partnering with. So it's quite an interesting time. The other thing that might be notable for your audiences, as a result of all of this innovation and disruption, we're also seeing the big players want to invest in and partner with not just the model makers, but with other application and enabling companies in the GenAI ecosystem. And not just an investment, but actually often interesting go-to-market distribution partnerships and their eagerness, these bigger cloud service providers and big software companies to do that is as strong as I've ever seen in my venture career.

Interesting. So we are almost at the end of our conversation and I ask all my guests at the end a couple of short, rapid questions. So the first one is, who are the mentors who shaped your career?

Well, there's, there's a couple mentors, and the one in my venture career specifically is really Tom Allberg. And I would also give a shout out to Paul Goodrich. Tom and Paul were two of the co-founders of, of Madrona. They've been great mentors to me on just how do you create an incredible lasting culture as a smaller firm like ours. But how do you identify incredible entrepreneurs and work alongside them to build companies? And then just, you know, just their example as people and human beings and how they acted with humility and integrity all their lives. Tom, of course, he passed away last year, but was just an absolutely amazing mentor. And then earlier in my career, I think back to people at, you know, McKinsey and investment banking and in my operating role, Larry Prince, who was the CEO of Genuine Parts at the time and Tom Gallagher and these folks that took an interest in you and were willing to, you know, both encourage you when you had a tough day and needed encouragement and also give you a little curbside coaching when, you know, I could do something better. And so, you know, if you don't have those folks willing to do that, you're not gonna develop and grow in your career.

What do you know about early stage investing now that you wish you know when you first started with Medtronic?

That's a great question on what I wish I knew. There's so much that I've learned over the last 25 years, you know, and I think, you know, we've got this triangle which we call sort of the founder market fit triangle. And it's about, you know, there's a problem and it needs to be solved and the founders need to be obsessed about solving it. There's a bunch of macro reasons, some are technological, some are societal about the why now, why could you solve this problem better now? And then there's sort of a, you know, what's going to be your solution or the why this question. But in the middle of that triangle is actually the most important, and that's of course the founders. And so since it's the founder market fit triangle, it's the why us. And listening carefully of why does that founding team believe that they can solve that problem? Why do they understand the context of the time? And do they have that combination of curiosity and humility and determination that it's going to take over a long period of time to go be the winners in building not only a product but a company around that. And so I think it's specifically thinking deeply about founders and their fit with an important problem that they're trying to solve. And I think that's something that you always are learning and getting better at. But, you know, the thing I would say is if you kind of ask that question in the context of the younger members of our team and other people in the investing space, you know, what to focus on.

What are your favorite books and anything interesting that you're reading?

Well, I absolutely, you know, love reading. I'm a slow reader, so I have to, you know, try to find mechanisms. I've become even more avid on the podcasting front over the last few years, something I really picked up during COVID But a couple of books that I would pick out that I think are just terrific books. So one is a company, a book called, sorry, Essentialism, which I can summarize as do less but better, which seems to fit because the book's called Essentialism. Just a terrific book. Then there's a book that Annie Duke wrote called Thinking in Bets. And this is about sort of probabilistic decisions. You never have perfect information. And she wrote a very engaging book, well applied to the business world. And then most recently, this book From Strength to Strength, which is Arthur Brooks, absolutely fantastic book. All about how you have certain strengths earlier in your career and other strengths later in your career. And if you can't make the leap of moving from the early strengths to the later strengths, not only are you going to be probably less successful later on in life, you're going to be less happy. And so focusing on the things that are going to be produced enduring happiness because you move from strength to strength. And he's a terrific author, actually a Seattle person originally. And so that's another book I'd I just love.

You mentioned podcasts. What are your favorite podcasts to listen to?

Oh, I have a number of podcasts I like. I do give a couple of shoutouts here. I really like Patrick O'Shaughnessy's Invest Like the Best. I think it's a very thoughtful, you know, podcast, a nice mix of investor types as well as entrepreneurs and innovators. I think, you know, it's really fun that you're probably familiar with Acquired. Ben and David actually started, acquired at Madrona many years ago. They're great friends of the firm, incredibly engaging and deep in the work that they do. And so I think it's distinctive, you know, from that perspective. And I would be remiss if I didn't mention The Room podcast, which is actually my daughter's podcast, which is all about people being in the room as innovators and entrepreneurs and what was their story of how they got in the room regardless of their background and circumstances and made something really special of it. And so I think that's just a fantastic podcast as well.

That's an interesting angle. I'll definitely check it out. What is your advice to any recent college grads that are interested in, you know, joining a venture firm?

Well, I think on the joining a venture firm or the founding a company for that matter, you know, this, this, you know, founder market fit triangle is a really important thing to, to think about. And if it's in the founding a company, it's especially important because there's going to be a lot of interesting ideas that you should have a high bar to decide, am I going to, you know, make that decision to put everything in? I'm basically all in, as it were, into starting a company, building a company, taking the risk. I think the other, you know, piece of advice is you've got to follow the people and the passions that are important to you. And, you know, so a lot of times people want to follow the thing that looks appealing at the time, but it's not the thing they're passionate about. or, you know, some people that have got, you know, kind of a, maybe a sort of a broader public appeal, but maybe they're not people they connect with as well. And so I think if you make those choices in favor of people that you really align with, you connect with, you can learn a lot from, they're gonna take an interest in you, the things that you are deeply passionate about, you're more likely to have the kind of career success, um, that, uh, that you have the potential to achieve.

Matt, this has been a very fascinating conversation. Thanks for joining us.

Well, thank you very much for having me. I really enjoyed it.

Mm-hmm.